KENO-NR: A Monte Carlo code simulating the californium-252-source-driven noise analysis experimental method for determining subcriticality.

Abstract

The $\sp{252}$Cf-source-driven noise analysis (CSDNA) requires the measurement of the cross power spectral density (CPSD) $G\sb{23}(\omega)$, between a pair of neutron detectors (subscripts 2 and 3) located in or near the fissile assembly, and the CPSDs, $G\sb{12}(\omega)$ and $G\sb{13}(\omega)$, between the neutron detectors and an ionization chamber (subscript 1) containing $\sp{252}$Cf also located in or near the fissile assembly. The key advantage of this method is that the subcriticality of the assembly can be obtained from the ratio of spectral densities,$${G\sbsp{12}{*}(\omega)G\sb{13}(\omega)\over G\sb{11}(\omega)G\sb{23}(\omega)},$$using a point kinetic model formulation which is independent of the detector's properties and a reference measurement. The multigroup, Monte Carlo code, KENO-NR, was developed to eliminate the dependence of the measurement on the point kinetic formulation. This code utilizes time dependent, analog neutron tracking to simulate the experimental method, in addition to the underlying nuclear physics, as closely as possible. From a direct comparison of simulated and measured data, the calculational model and cross sections are validated for the calculation, and KENO-NR can then be rerun to provide a distributed source $k\sb{eff}$ calculation. Depending on the fissile assembly, a few hours to a couple of days of computation time are needed for a typical simulation executed on a desktop workstation. In this work, KENO-NR demonstrated the ability to accurately estimate the measured ratio of spectral densities from experiments using capture detectors performed on uranium metal cylinders, a cylindrical tank filled with aqueous uranyl nitrate, and arrays of safe storage bottles filled with uranyl nitrate. Good agreement was also seen between simulated and measured values of the prompt neutron decay constant from the fitted CPSDs. Poor agreement was seen between simulated and measured results using composite $\sp6$Li-glass-plastic scintillators at large subcriticalities for the tank of uranyl nitrate. It is believed that the response of these detectors is not well known and is incorrectly modeled in KENO-NR. In addition to these tests, several benchmark calculations were also performed to provide insight into the properties of the point kinetic formulation.